Module 4: Multispectral Data Interpretation

    In this week’s lab assignment, I learned about techniques that are used to identify features in ERDAS and ArcGIS Pro using multispectral imagery. Then, I utilized those techniques to identify features based on their spectral signature.

    In exercise 1, I explored accessing and downloading Landsat satellite imagery from Glovis. Then in exercise 2, I utilized high pass and low pass filters in ERDAS to understand detailed and broad patterns in imagery. Next I learned about application of the focal statistics tool in ArcGIS Pro and compared output from ERDAS and ArcGIS Pro.

    In exercise 3, I accessed the histogram of imagery in ERDAS and ArcGIS Pro to look at the distribution of pixel values with a layer of multispectral imagery. I learned how to interpret the histogram graphic and alter breakpoints in both programs. Continuing to Exercise 4, I explored how altering the band combinations affected the visibility of certain features. In particular, I was interested in understanding what band combinations made logging roads more visible. I personally found that the TM False Natural Color band Combination (Red:5, Green: 4, and Blue: 3) made the roads standout the best. The roads appeared pink which allowed them to stand out against the largely green background (forest).

    For exercise 5, I used the NDVI tool to calculate the normalized differential vegetation index (NDVI) using multispectral imagery of Washington state. Then, I compared the NDVI layer values for different features in the image. Next, in exercise 6 I utilized the metadata to identify important information about the layers such as the data type, maximum/minimum pixel values, and the pixel size (image resolution). Understanding this information helps me make inferences about the broad scale trends in an image (e.g. Does the image feature mostly urban areas? forested areas? water?).

    Last, in exercise 7 I applied many of the skills learned in the previous lab exercises and to identify features based on their spectral signatures. Altering the multispectral bands and knowing what to look for in the output based on histogram plots is an extremely valuable skill set. Using band combinations that I did not necessarily explore in the previous lab exercises helped me create maps that highlighted the features of interest more clearly. Below is one of the maps that I made for the assignment. By using a custom band combination (R: band 2, G: band 3, B: band 6), the snow caps are highlighted as a bright yellow which sharply juxtaposes the blue/purple features in the background.